"Analysis of Clinical Trials Using SAS: A Practical Guide"
Alex Dmitrienko, Geert Molenberghs, Christy Chuang-Stein, Walt Offen

If you would like to share comments or feedback, we encourage you to contribute to this discussion thread by clicking on the Post your reply or comments link at the top of the page.

For more information on this SAS Press book, please visit its web page at

http://www.biopharmnet.com/books/book40005.html

The following changes need to be made in the first printing (books printed in 2005, 2006 and early 2007):

Page 25, Equation (1.9) needs to be corrected. Delete "hbox".

Page 125, Several changes need to be made on this page:
Replace "\delta_{ij}(S)/v_{ij}(S)" with "\delta_{ij}(S)p_{ij}/v_{ij}(S)".
Step 2, Replace "v_{1}^{*}" with "v_{2}^{*}".

Page 129, Third line from the bottom
"Dmitrienko et al (2004)" needs to be replaced with "Dmitrienko et al (2005)"

Page 131, Paragraph after Program 3.1
"Dmitrienko et al (2004)" needs to be replaced with "Dmitrienko et al (2005)"

Page 133, Program 3.3, Line 7
"%macro DistPlot(dataset);" needs to be replaced with "%macro DistPlot(dataset,label);"

Page 220, Program 4.9, Line 4
"%EffMonitor(fraction=fraction,data=stat,effsize=0.375,power=0.9,"
needs to be replaced with
"%EffMonitor(fraction=SepTrial,data=stat,effsize=0.375,power=0.9,"

Page 223, Program 4.10, Line 4
"%EffMonitor(fraction=fraction,data=stat,effsize=0.375,power=0.9,"
needs to be replaced with
"%EffMonitor(fraction=SepTrial,data=stat,effsize=0.375,power=0.9,"

Page 234, Line 1 in Section 4.2.9
Replace "Sequential designs have long become" with "Sequential designs have become"

Page 236, Paragraph 2
"Dmitrienko and Wang (2004)" needs to be replaced with "Dmitrienko and Wang (2006)"

Page 251, Line 1
"Dmitrienko and Wang (2004)" needs to be replaced with "Dmitrienko and Wang (2006)"

Page 257, Line 3
"Dmitrienko and Wang (2004)" needs to be replaced with "Dmitrienko and Wang (2006)"

Page 261, Bottom line
"Dmitrienko and Wang (2004)" needs to be replaced with "Dmitrienko and Wang (2006)"

Page 264, Paragraph 4
"Dmitrienko and Wang (2004)" needs to be replaced with "Dmitrienko and Wang (2006)"


Page 393
Replace
"Dmitrienko, A., Sides, G., Winters, K., Kovacs, R., Eisenberg, P., Groh, W.
(2004). Electrocardiogram reference ranges derived from a standardized clinical trial population. Journal of Cardiovascular Electrophysiology. Forthcoming."
with
"Dmitrienko, A., Sides, G., Winters, K., Kovacs, R., Eisenberg, P., Groh, W. (2005). Electrocardiogram reference ranges derived from a standardized clinical trial population. Drug Information Journal. 39, 395-406."

Page 393
Replace
"Dmitrienko, A., Wang, M.D. (2004). Bayesian predictive approach to interim monitoring in clinical trials. Unpublished manuscript."
with
"Dmitrienko, A., Wang, M.D. (2006). Bayesian predictive approach to interim monitoring in clinical trials. Statistics in Medicine. 25, 2178-2195."

Page 410, K
"Kaplan-Neier estimates" needs to be replaced with "Kaplan-Meier estimates"

Last edited by BioPharmNet on Fri Feb 01, 2008 5:18 pm; edited 1 time in total

Yun Peng's question:

I am trying to use the minimum risk model macro in your book 'Analysis of Clinical Trials Using SAS'. I'd like to know in case I have one strata with p1=p2=1, what should I do? My V matrix will have an entry of 0. Does adding a very small number (for example 0.00000001) to V to make the zero be able to take the inverse make sense? But then the number of
observations contributing to this p1=p2=0 will not be reflected. Do you have any suggestions?

Devan Mehrotra's response:

Let X_i/N_i denote the sample proportion for treatment i. To avoid a zero variance that occurs when both X_i=0, replace each X_i with X_i+0.5. Likewise, to avoid a zero variance that occurs when both X_i=N_i, replace each X_i with X_i-0.5. Note that if only one of the sample proportions is zero (or one), then no adjustment is necessary for any of the X_i's because the variance will not be zero.

Extensive simulations over the years have shown that the above strategy works fine, and I recommend it in the following publication:
Mehrotra, DV. (2001). Stratification Issues with Binary Endpoints. Drug Information Journal. 35(4):1343-1350.

The MinRisk macro in Section 1.3.3 does not currently generate a confidence interval for this method. For that, you will need to use the observed rather than the null variance. In my latest research (see part I of the attached PowerPoint slides), I recommend using the minimum risk weights along with the observed rather than the null variance, for both superiority and non-inferiority trials.

PowerPoint slides
http://www.biopharmnet.com/doc/2008_01_10_presentation.ppt